Coronary artery disease is the most common cause of death world­wide. On­pump coronary artery bypass grafting and cardioplegia remains the gold standard for the treatment of multivessel coronary disease. However, this method of surgical treatment has a number of perioperative complications, the most common of which is post­operative cognitive dysfunction (POCD).Aim. To develop a mathematical model for predicting the binary outcome "presence/absence of POCD deterioration" using Mini­Mental State Examination (MMSE) and Frontal Assessment Battery (FAB) for timely prevention of cognitive impairment in the early postoperative period.Material and methods. The study included 180 patients with coronary artery disease with stable angina who underwent on­pump coronary artery bypass grafting. The patients’ age [Me (Q25­-Q75)] was 62 [56­-67] years. After the operation, all patients were divided into 2 groups: 108 patients without POCD and 72 patients with cognitive deficit of various severity. The study was conducted in three stages: stage I — 2 days before surgery; stage II — during the operation; stage III — 7 days after the operation. Assessment of cognitive status using FAB, MMSE tests was carried out at stages I and III. At the second stage, biochemical parameters were assessed: in arterial blood — lactate and hemoglobin levels, in venous blood — pH, partial pressure of carbon dioxide (pCO2, mm Hg) and partial pressure of oxygen (pO2, mm Hg), as well as the duration of operation, cardiopulmonary bypass and myocardial anoxia. Discriminant analysis was used to develop a mathematical model for predicting a binary outcome.Results. Two groups of the most informative indicators were identified that were included in prediction algorithms for binary outcomes "presence/absence of negative dynamics of POCD" for MMSE and FAB. Group 1 included perioperative MMSE values, atrial fibrillation before surgery, cardiopulmonary bypass time and pO2. Group 2 included cardiopulmonary bypass time and intraoperative venous pH. Conclusion. Mathematical models have been developed for predicting the binary outcome "presence/absence of POCD deterioration" ac­cording to the MMSE and FAB cognitive tests, which make it possible to assess the possibility of postoperative cognitive dysfunction. The proposed algorithms are implemented using spreadsheets and a computer program.